Verified AI Contributions: What They Are and How They Work

A verified AI contribution is a prompt that is cryptographically linked to a real git commit. The commit SHA exists on GitHub, GitLab, or Bitbucket. The prompt timestamp falls within the commit window. The files referenced in the prompt match files changed in the commit. This creates a chain of evidence that is independently verifiable.

How Verification Works

When you commit code, Qmmit captures your AI prompts and stores them locally with the commit SHA. When you push, the prompts are synced to the platform along with the commit metadata. The platform then verifies: does this commit SHA exist on the linked repository? Does the timestamp make sense? Do the files match?

For public repositories, anyone can verify a contribution by checking the commit SHA on GitHub. The prompt timestamp, model name, and matched files are all visible on the developer profile.

Confidence Levels

Not all matches are equal. Qmmit assigns confidence levels based on the matching algorithm score. HIGH (80%+) means the prompt is strongly linked to the commit — the timing, files, and content all align. MEDIUM (50-79%) means there is a reasonable match but some ambiguity. LOW (below 50%) means the match is weak and needs manual review.

High-confidence matches are displayed prominently on profiles. Medium matches are shown with a note. Low matches go to a review queue where the developer can confirm or reject them.

What Makes This Different from Self-Reporting

Self-reported AI usage (like adding "Built with AI" to a README) has no verification. Anyone can claim anything. Verified contributions are backed by git commit SHAs, timestamps, and file-level matching. The evidence is independently checkable.

Cross-Platform Support

Verification works across GitHub, GitLab, and Bitbucket. Qmmit auto-detects your git platform from .git/config and uses the appropriate API to verify commits. Self-hosted instances (GitHub Enterprise, GitLab CE/EE, Bitbucket Server) are also supported.

Building Trust

The goal of verified AI contributions is to build trust in the developer ecosystem. When a recruiter sees a profile with 500 verified prompts across 10 projects, they know those contributions are real. When a maintainer reviews a PR from a developer with a strong AI attribution history, they have more context for the review.